As the world is moving towards paperless exchange of information, use of documents and document repositories including such documents is on the rise. Currently, thumbnail of a document does not convey human-readable information about the specifics of the document unless a user of the document clicks the thumbnail to open the document. Further, in case of similar type of documents the thumbnails are similar looking and do not convey any meaningful information to users. Hence, a user of a document repository has to access each document, individually, to obtain the specifics of each document. The situation worsens in cases in which the documents are text-heavy documents and do not include any images. Therefore, there is a need for improved way of generating thumbnails.